Data packaging tools for the HEAL data ecosystem
Project description
--8<-- [start:intro]
HEAL Data Utilities
The HEAL Data Utilities python package provides data packaging tools for the HEAL Data Ecosystem to facilitate data discovery, sharing, and harmonization with a focus on the HEAL Platform Data Consultancy (DSC).
Currently, the focus of this repository is generating data dictionaries (see Variable-level Metadata section below). However, in the future, this will be expanded for all HEAL-specific data packaging functions (e.g., study- and file-level metadata and data).
Installation
To install the latest official release of healdata-utils, from your computer's command prompt, run:
pip install healdata-utils --pre
(NOTE: currently in pre-lease)
pip install git+https://github.com/norc-heal/healdata-utils.git
--8<-- [end:intro]
--8<-- [start:vlmd-intro]
Variable-level Metadata (Data Dictionaries)
The healdata-utils variable-level metadata (vlmd) tool inputs a variety of different input file types and exports HEAL-compliant data dictionaries (JSON and CSV formats). Additionally, exported validation (i.e., "error") reports provide the user information as to a) if the exported data dictionary is valid according to HEAL specifications and b) how to modify one's data dictionary to make it HEAL-compliant.
--8<-- [end:vlmd-intro]
--8<-- [start:vlmd-basic-usage]
Basic usage
The vlmd tool can be used via python or the command line.
Using from python
From your current working directory in python, run:
from healdata_utils.cli import convert_to_vlmd
# description and title are optional. If submitting through platform, can fill these out there.
description = "This is a proof of concept to demonstrate the healdata-utils functionality"
title = "Healdata-utils Demonstration Data Dictionary"
healdir = "output" # can also specify a file name if desired (eg output/thisismynewdd.csv)
inputpath = "input/my-redcap-data-dictionary-export.csv"
data_dictionaries = convert_to_vlmd(
filepath=inputpath,
outputdir=healdir,
inputtype=input_type, #if not specified, looks for suffix
data_dictionary_props={"title":title,"description":description} #data_dictionary_props is optional
)
This will output the data dictionaries to the specified output directory (see output section below) and also save the json/csv versions in the
data_dictionaries
object.
For the available input file formats (i.e., the available choices for the
inputtype
parameter), one can run (from python):
from healdata_utils.cli import input_descriptions
input_descriptions
The input_descriptions
object contains the choice for inputtype
as the key and the description as the value.
Using from the command line
From your current working directory run:
(note the \
at the end of each line signals a line continuation for ease in understanding the long, one-line command.) Again, the --title
and --description
options are optional.
For descriptions on the different flags/options, run vlmd --help
vlmd --filepath "data/example_pyreadstat_output.sav" \
--outputdir "output-cli" \
--title "Healdata-utils Demonstration Data Dictionary" \
--description "This is a proof of concept to demonstrate the healdata-utils functionality"
Output
Both the python and command line routes will result in a JSON and CSV version of the HEAL data dictionary in the output folder along with the validation reports in the errors
folder. See below:
-
input/input/my-redcap-data-dictionary-export.csv
: your input file -
output/errors/heal-csv-errors.json
: outputted validation report for table in csv file against frictionless schema- see schema here
-
output/errors/heal-json-errors.json
: outputted jsonschema validation report.- see schema here
!!! important The main difference* between the CSV and JSON data dictionary validation lies in the way the data dictionaries are structured and the additional metadata included in the JSON data dictionary.
The CSV data dictionary is a plain tabular representation with no additional metadata, while the JSON dataset includes fields along with additional metadata in the form of a root description and title.
* for field-specific differences, see the schemas in the documentation.
output/heal-csvtemplate-data-dictionary.csv
: This is the CSV data dictionaryoutput/heal-jsontemplate-data-dictionary.json
: This is the JSON version of the data dictionary
Note, only the JSON version will have the user-specified
title
anddescription
Interactive notebooks
See the below notebooks demonstrating use and workflows using the convert_to_vlmd
in python and vlmd
in the command line.
Clicking on the "binder badges" will bring you to an interactive notebook page where you can test out the notebooks. Here, healdata-utils comes pre-installed.
- Generating a heal data dictionary from a variety of input files
- click here for static notebook
- click binder badge for interactive
- [in development] Creating and iterating over a csv data dictionary to create a valid data dictionary file click here
--8<-- [end:vlmd-basic-usage]
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for healdata_utils-0.1.6a0-py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | dcd20a0e4cb8f0971117094c4a8cb7a0262652c7bc525281adca703e572e922e |
|
MD5 | 37f2fd4a5d8a8f8ecddc00fab45b2895 |
|
BLAKE2b-256 | 29ebcc532ab8dfe0f1696ba783b1d8397e05db473cb22101ccaa4248e489163e |